Re: periodic boundary condition in XY, hard wall in Z - errors

From: Olya Kravchenko (ovkrav_at_gmail.com)
Date: Thu Mar 10 2016 - 20:01:12 CST

Thank you! I will study MSM method.

I have changed the size of cell vectors back to 62, that is the actual
size rounded up, and I commented out cellBasisVector3. I ran a number
of simulations today and I keep seeing the same error when I increase
the number of timesteps. The margin size does seem to make the job run
longer, at least I don't see any other correlation. My last run quit
near 150 000 timesteps (I set it up for 500K timesteps).

Apart from that, I see in vmd that my simulation box actually becomes
longer in Z-direction, it is a cube in the beginning and then the
proportion changes. I am not sure how to interpret it, but it looks
like the size in z does not hold, certainly no hard wall. I wonder if
there is a way to set up the hard wall in a simple way?

Olga

On Thu, Mar 10, 2016 at 4:58 PM, Maxim Belkin <mbelkin_at_ks.uiuc.edu> wrote:
> Olya,
>
> If your system size is 60A in x and y, you should use vectors of length 60 as your "cellbasisvector[1,2]". You are currently using 70, which means you have vacuum in x and y directions and this is not what you want. To simulate a system that is non-periodic along Z with NAMD you have to use Multi-level summation method (look for MSM option in NAMD User’s Guide). I believe you have to remove "cellbasisvector3" instead of specifying 0 0 0 but I’m not 100% positive on that.
>
> Maxim
>
>
>> On Mar 10, 2016, at 11:21, Olya Kravchenko <ovkrav_at_gmail.com> wrote:
>>
>> My system consists of 200 argon atoms in a box, PBC in X and Y, I
>> would like hard wall in Z and my goal is to measure
>> density/concentration. So I think that NPT is reasonable in my case,
>> is that correct?
>>
>> I commented out PME, because I read that it will speed up
>> calculations, also because, if I understand it correctly, it requires
>> all three cell basis vectors to be defined.
>>
>> linaccel is commented because I defined my own force in a separate
>> file for the force.
>>
>> Here is my configuration file:
>>
>> #############################################################
>> ## JOB DESCRIPTION ##
>> #############################################################
>>
>> # Equilibration of argon atoms with PBC in XY direction
>>
>> #############################################################
>> ## ADJUSTABLE PARAMETERS ##
>> #############################################################
>>
>> structure ar.psf
>> coordinates ar.pdb
>>
>> set temperature 310
>> set outputname ar_force_out
>>
>> firsttimestep 0
>>
>>
>> #############################################################
>> ## SIMULATION PARAMETERS ##
>> #############################################################
>>
>> # Input
>> paraTypeCharmm on
>> parameters My_parameters.rtf
>> temperature $temperature
>>
>>
>> # Force-Field Parameters
>> exclude scaled1-4
>> 1-4scaling 1.0
>> cutoff 12.0
>> switching on
>> switchdist 10.0
>> pairlistdist 14.0
>>
>>
>> # Integrator Parameters
>> timestep 2.0 ;# 2fs/step
>> rigidBonds all ;# needed for 2fs steps
>> nonbondedFreq 1
>> fullElectFrequency 2
>> stepspercycle 10
>>
>>
>> # Constant Temperature Control
>> langevin on ;# do langevin dynamics
>> langevinDamping 1 ;# damping coefficient (gamma) of 1/ps
>> langevinTemp $temperature
>> langevinHydrogen off ;# don't couple langevin bath to hydrogens
>>
>>
>> # Periodic Boundary Conditions
>> cellBasisVector1 62.0 0. 0.0
>> cellBasisVector2 0.0 62.0 0.0
>> cellBasisVector3 0.0 0 0.0
>> cellOrigin 30.0 30.0 32.5
>>
>> wrapAll on
>>
>> margin 10
>>
>> # PME (for full-system periodic electrostatics)
>> #PME yes
>> #PMEGridSpacing 1.0
>>
>>
>> # Constant Pressure Control (variable volume)
>> useGroupPressure yes ;# needed for rigidBonds
>> useFlexibleCell no
>> useConstantArea no
>>
>> langevinPiston on
>> langevinPistonTarget 1.01325 ;# in bar -> 1 atm
>> langevinPistonPeriod 100.0
>> langevinPistonDecay 50.0
>> langevinPistonTemp $temperature
>>
>> ##################
>> ###DEFINE FORCES##
>> ##################
>>
>> tclForces on
>> tclForcesScript my_force.tcl
>> #set linaccel "30 0 0"
>>
>> # Output
>> outputName $outputname
>>
>> restartfreq 500 ;# 500steps = every 1ps
>> dcdfreq 250
>> xstFreq 250
>> outputEnergies 100
>> outputPressure 100
>>
>>
>> # Minimization
>> minimize 100
>> reinitvels $temperature
>>
>> run 100000
>>
>> Thanks!!
>>
>> Olga
>>
>> On Thu, Mar 10, 2016 at 2:05 AM, Norman Geist
>> <norman.geist_at_uni-greifswald.de> wrote:
>>> We need to see your full input script, since it looks you're doing npt?
>>>
>>> Norman Geist
>>>
>>>> -----Ursprüngliche Nachricht-----
>>>> Von: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] Im
>>>> Auftrag von Olya Kravchenko
>>>> Gesendet: Donnerstag, 10. März 2016 04:23
>>>> An: NAMD <namd-l_at_ks.uiuc.edu>
>>>> Betreff: namd-l: periodic boundary condition in XY, hard wall in Z - errors
>>>>
>>>> Hi all,
>>>>
>>>> I would like to apply periodic boundary conditions only in X and Y
>>>> directions. In my config file I used the following entry:
>>>>
>>>> # Periodic Boundary Conditions
>>>> cellBasisVector1 70.0 0. 0.0
>>>> cellBasisVector2 0.0 70.0 0.0
>>>> cellBasisVector3 0.0 0 0.0
>>>> cellOrigin 30.0 30.0 32.5
>>>>
>>>> wrapAll on
>>>> margin 5
>>>>
>>>>
>>>> When I submit the job it runs for a while until I see the following error:
>>>>
>>>> WRITING COORDINATES TO DCD FILE ar_force_out.dcd AT STEP 20750
>>>> FATAL ERROR: Periodic cell has become too small for original patch grid!
>>>> Possible solutions are to restart from a recent checkpoint,
>>>> increase margin, or disable useFlexibleCell for liquid simulation.
>>>>
>>>> How can I address this error? Is this the right way to set up PBC in 2
>>>> dimensions?
>>>>
>>>> My simulation cell is actually around (~61 ~61 ~62), as measured by
>>>> minmax. I played with numbers and noticed that my simulation runs
>>>> longer if I increase both the margin and cell basis vectors.
>>>>
>>>> I saw some of the older threads in the archives where the margin was
>>>> discussed but it's not clear how large should it be. If I want to run
>>>> my simulation for 500 000 steps, what margin should I set up?
>>>>
>>>> Also, when choosing cell basis vector, how much bigger should it be
>>>> than what is measured by "measure minmax"?
>>>
>

This archive was generated by hypermail 2.1.6 : Tue Dec 27 2016 - 23:21:51 CST